TEXT

Top Abstract Text References

Yersinia outer proteins (Yops) are virulence determinants synthesized by the Yersinia species pathogenic for humans, including Y. pestis, the causative agent of plague (29). The Yop proteins are encoded on a 75-kb plasmid, and in vitro expression from these genes and subsequent secretion and translocation by a Type III secretion system are regulated by temperature, calcium, and eukaryotic cell contact (11, 13, 18, 32, 34). There are various functions known for some of the Yops, including a cytotoxic effect on the host cell, with depolymerization of actin microfilaments (YopE [31, 32]); dephosphorylation of host cell proteins (YopH [2, 6]); serine/threonine kinase activity (YpkA [16, 17, 20]); translocation and sensor functions (YopB/D and YopN [7, 14, 19-21, 28]); inhibition of platelet aggregation (YopM [23, 30]); and control of Yop translocation, as well as establishment of a systemic infection (YopK [22, 23, 35]).

Previous studies showed that antibodies to some Yops are present in convalescent-phase serum from patients infected with Y. pestis, as well as in rodent serum after experimental Y. pestis infection (5, 8, 12, 25, 27), which suggests that Yops are antigenic. Furthermore, vaccination with Yop-containing culture supernatants from growth-restricted Yersinia enterocolitica protected mice from a lethal intraperitoneal (i.p.) dose of virulent Y. pestis (26). Interpretation of the latter results, however, is complicated by the likely presence of V antigen in the crude supernatants, as V is known to be a protective antigen (1).

To define further the role of individual Yops in plague immunity, we evaluated six recombinant Y. pestis Yop products, YopD, YopE, YopH, YopK, YopN, and YpkA, for their ability to elicit a protective immune response against subcutaneous (s.c.) plague infection in an animal model. Wild-type Y. pestis, CO92, was obtained from the laboratory of T. Quan, Centers for Disease Control, Ft. Collins, Colo., and an attenuated pigmentation-negative (Pgm) derivative of this strain was isolated by serial passage on Congo red medium (36). The variant was used to obtain template plasmid DNA for amplifying yopD, E, H, K, and N and ypkA by PCR. The specific sequences used for amino- (11 to 13 bases 5' to 3' from the ATG start codon) and carboxy- (11 to 13 bases 3' to 5' from the stop codon) terminal oligonucleotide primers for the yop open reading frames (ORFs) were identical to those sequences published for Y. pseudotuberculosis or Y. pestis (9, 14-17, 19, 23). PCRs were performed with purified plasmid DNA combined with each primer pair by use of the Perkin Elmer GeneAmp reagent kit with AmpliTaq polymerase (Roche Molecular Systems, Inc., Branchburg, N.J.) according to the manufacturer's procedures. The PCR products were ligated in frame into the expression vector pET19b (Novagen, Inc., Madison, Wis.). The resulting recombinant plasmids encoded gene fusions with a 10-histidine amino terminus followed by a porcine enterokinase site immediately upstream of the ATG start codon of the respective yop ORF. Standard genetic manipulations (33) were used to ligate, transform, and select for positive recombinant clones in Escherichia coli DH5- (Bethesda Research Laboratories, Gaithersburg, Md.). Recombinant plasmids, containing yop inserts, were then transferred into the protease-deficient (ompT lon) E. coli strain BL21(DE3) or BL21(DE3)pLysS (Novagen), which contains a copy of the T7 polymerase gene under the control of an isopropyl--D-thiogalactopyranoside (IPTG)-inducible promoter.

DNA of the yop gene fusion isolates was sequenced with purified recombinant plasmid double-stranded template DNA by use of a Sequenase version 2.0 DNA sequencing kit (U.S. Biochemical Corp., Cleveland, Ohio) following the instructions of the manufacturer. Sequencing analysis of positive yopD, yopE, yopK, and ypkA fusions revealed that the sequences of the cloned PCR products were identical to published sequences for the four genes.

Y. pestis yopH and yopN loci were PCR amplified, cloned into pET21a (Novagen), and sequenced under contract by PerImmune, Inc, Rockville, Md. The sequences were identical to published sequences except for a silent mutation in yopH, in which a T was replaced by a C at base 1237 of the ORF. These constructs expressed full-length recombinant products without a histidine tag (YopH and YopN) from E. coli BL21(DE3)pLysS when induced with 1 mM IPTG.

Histidine-tagged YopD, YopE, YopK, and YpkA from recombinant strains were expressed and purified by IPTG induction of batch cultures and Ni²⁺ chelation chromatography under denaturing conditions (6 M urea) according to the procedures specified by the manufacturer (Novagen). YopH was purified by high-performance liquid chromatography (HPLC) cation exchange and size-exclusion chromatography under nondenaturing conditions, and YopN was purified by HPLC anion exchange and size-exclusion chromatography, after solubilization in 6 M urea (PerImmune, Inc.). The urea-extracted proteins were step dialyzed against 3, 1, and 0.5 M urea and phosphate-buffered saline.

Shown in Figure 1 are samples of post-chromatographed, dialyzed YpkA, YopD, YopE, YopH, YopK, and YopN run on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and stained with Coomassie blue (33). The proteins were greater than 90% pure, with only minor contaminating species present. The molecular masses of YpkA (85 kDa), YopD (36 kDa), YopE (26 kDa), and YopK (21 kDa) agreed with the predicted molecular masses of the histidine-tagged recombinant species. Additionally, the masses of the YopH and YopN products (50 and 27 kDa, respectively) matched the masses of the respective species predicted by the sequence.

View larger version (48K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE analysis of purified recombinant Yops. Lanes: 1, YpkA; 2, YopD; 3, YopE; 4, YopH; 5, YopK; 6, YopN. The positions of molecular mass standards (in kilodaltons) are indicated on the left axis.

For vaccination, the purified proteins were mixed with an equal volume of the adjuvant R-730 emulsion (Ribi ImmunoChem Research, Inc., Hamilton, Mont.) to give a final protein concentration of 100 µg/ml. Thirty micrograms of each Yop-adjuvant mixture was next administered s.c. to two groups each of 8 to 14 female, 8-week-old, Hsd:ND4 Swiss Webster outbred mice (Harlan Sprague Dawley, Indianapolis, Ind.), followed by one boost s.c. of 30 µg at 30 days post-primary vaccination. Serum samples from each Yop-vaccinated mouse were analyzed for antibody to individual Yops by an enzyme-linked immunosorbent assay (ELISA) with urea-solubilized Yops as the coating antigens on glutaraldehyde-pretreated microtiter plates (24). After two doses, all Yops except YopE induced high antibody titers in vaccinated mice (data not shown). Except as indicated, mice in all groups were subsequently boosted with 30 µg of each antigen (15 µg s.c. and 15 µg i.p.) on day 60. The YopE mice were boosted once more s.c. and i.p. on day 90. In the YpkA experiment and in a second YopD experiment (Tables 1 and 2), mice were vaccinated with only two doses. The mice were then retested for Yop antibody titers. A control group was vaccinated with the adjuvant R-730 emulsion alone. Mice in additional control groups were each given 0.1 ml of the killed whole-cell plague vaccine licensed for human use (Plague Vaccine USP, lot number 10K01A; Miles, Inc.; Cutter Biological, Elkhart, Ind.) or recombinant Y. pestis F1 capsular antigen in R-730 emulsion (10 µg/dose) in a three-dose s.c./i.p. vaccination regimen. The recombinant F1 capsular antigen was produced from E. coli and purified from cells as described previously (3). During all vaccinations and subsequent challenges, animals were provided with fresh water and feed ad libitum. Additionally, all experiments were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals (10).

After the three- to four-dose vaccination regimen, all Yop-vaccinated animal groups possessed high titers of antibody to the recombinant proteins, while mice in the control group, given R-730, possessed negligible anti-Yop antibody titers. The reciprocal geometric mean endpoint titers were greater than 275,000 for all the Yops except YopE, for which it was 50,686. Additionally, serum samples from the control animal group vaccinated with purified Y. pestis recombinant F1 capsular antigen in R-730 did not cross-react with any of the Yops in an ELISA (data not shown).

The vaccinated mice were challenged s.c. 4 weeks after the last dose with multiples of the dose lethal for 50% of nonvaccinated animals (LD₅₀) of either wild-type Y. pestis CO92 or a nonencapsulated, virulent isogenic derivative, C12 (38). The s.c. LD₅₀ is 1.9 CFU for CO92 (37) and 9.1 CFU for C12 (38). The organisms were grown and the CFU were measured as previously described (1).

The effectiveness of vaccination against the encapsulated Y. pestis CO92 presented in Table 1 was analyzed by overall survival at 28 days postchallenge by Fisher's exact test and by mean survival time by the log-rank statistic obtained from product-limit survival estimates from the BMDP Lifetest program (BMDP Software, University of California, Los Angeles). None of the Yops tested, including YopD (two experiments), YopH, YopK, YopM, and YopN, significantly protected mice against challenge with encapsulated CO92 (Table 1; P = 0.206 to 1.000). The only group showing any protection by the more sensitive indicator of mean survival time was YopD, and we observed this in only one of the two experiments (Table 1; P = 0.025 and P = 0.936, respectively). As previously described (3), both F1 capsular antigen itself and Plague Vaccine USP provided significant protection against death after CO92 challenge (Table 1; P < 0.001 and P = 0.002, respectively).

As expected, when F1- or Plague Vaccine USP-vaccinated animals were challenged with the F1-negative nonencapsulated C12 strain, there was no statistically significant protection observed (Table 2; P = 0.471 and P = 1.000, respectively). It is possible that the F1 capsule could mask some surface-exposed Yop proteins and thus prevent protection by antibodies to such antigens. If this was the case, then it should be easier to protect against capsule-negative strains than against capsule-positive strains. Table 2 shows that vaccination with YopH, YpkA, YopE, YopK, YopM, or YopN did not protect mice challenged with the nonencapsulated C12 strain against death (0 to 30% survival; P = 0.114 to 1.0). In these groups, only the YpkA-vaccinated mice had a significantly increased mean survival time, 14.1+/3.4 days, compared with controls (Table 2, P = 0.022). While YopD-vaccinated animals were not protected against challenge with encapsulated Y. pestis, as noted above, they were protected against challenge with the virulent, nonencapsulated C12, with a statistically significant increase in survival (Table 2; experiment 2, 90%, P = 0.001) and a significant increase in mean survival time (26.2+/1.7 days, P < 0.001) in one experiment. We observed protection against a higher-dose C12 challenge in a second experiment with 78% survival of YopD-vaccinated mice. While the increased survival observed in the second experiment did not reach statistical significance (Table 2; experiment 1, P = 0.057), the mean survival time was again highly significant (22.4+/3.1 days, P = 0.006).

On the basis of our results, humoral immunity to most of the individual Yops tested was not sufficient to afford good protection against Y. pestis infection introduced by the s.c. route. We make this conclusion with the caveat that four of the recombinant antigens which did not protect (YopE, YopK, YopN, and YpkA) were denatured before their use as immunogens. Thus, the manipulation of these proteins during extraction and purification may have destroyed their protective epitopes, if they exist. Some evidence, however, does suggest that some of the urea-solubilized recombinant Yops do retain some epitopes present on the native proteins. In preliminary experiments, mouse hyperimmune serum generated with recombinant YopD, YopE, YopH, and YopN, but not with YpkA, reacted with the respective native proteins by a whole-cell surface-labeling immunoassay technique (4) (data not shown). Conversely, in a separate study (5) some animals that survived experimental plague infection produced antibodies during the course of infection that reacted against urea-solubilized YopK with the same ELISA methodology described in the present report.

Of interest, vaccination with YopN, which is thought to function as a surface-located sensor for low-calcium- or eucaryotic-cell-contact-induced Yop secretion (7, 14), prolonged mean survival time only slightly after challenge with nonencapsulated C12 compared with the control group (Table 2; 8.0+/1.7 days versus 5.4+/0.4 days), but this difference did not reach statistical significance (P = 0.091). In regard to YpkA, vaccination did not affect overall survival but did result in a statistically significant increase in mean survival time of mice challenged with the nonencapsulated C12 strain (Table 2, P = 0.022). It is possible that YpkA has an even greater protective effect when purified in its native (active) form. However, as suggested by a model of vectorial translocation of the effector Yops, YpkA, as well as YopH and YopE, may remain intracellular in the bacteria until eucaryotic cell contact, after which they are delivered to the host cell cytosol (14, 24). Thus, it is possible that antibody to these Yops in the vaccinated host would not have access to the antigens. The failure of YpkA, YopE, and YopH antibody to protect mice against Y. pestis challenge in our experiments is consistent with this model.

A study by Nemeth and Straley (27) also showed that mice vaccinated with purified native Y. pestis YopM, an alpha-thrombin-binding protein which is also translocated to the host cell cytosol (7, 27), were not protected against an intravenous challenge with the Pgm Y. pestis strain KIM5. We obtained similar results when mice vaccinated with identical preparations of YopM (kindly provided by S. Straley, University of Kentucky, Louisville) were challenged s.c. with the fully virulent Y. pestis strains CO92 (Table 1) and C12 (Table 2). These results supported the hypothesis that some Yops may not be accessible for neutralization by antibody in the Y. pestis-infected host.

The ability of YopD to protect mice against nonencapsulated Y. pestis C12 strongly suggests that at least one of the Yops is important in eliciting a protective immune response against lethal Y. pestis s.c. challenge. Given the accessory function of YopD in facilitating delivery of other Yops into their eukaryotic targets, it is possible that antibody to YopD interferes with this function. The failure of YopD to protect against encapsulated organisms to the same degree as it protects against the nonencapsulated strain may be the result of a masking effect of the F1 capsule on secreted YopD which blocks the antibody-antigen interaction at the surface of the bacterium. Experiments are currently being conducted in our laboratory to examine these hypotheses.